[01:39] <fabioval> someone could direct me to a good tutorial on drbd + gfs?
[02:21] <james1980> where does dovecot store its error logs?
[02:27] <ikonia> james1980: /var/log - probably in /var/log/mail
[02:31] <james1980> both empty "new file"
[02:31] <ikonia> they are directories
[07:06] <Syria> hi
[07:07] <Syria> I have deleted apache but I am still getting the "It works" in my broswer !
[07:07] <Syria> This did not help as well sudo update-rc.d -f apache2 remove
[07:09] <Syria> Hi, I have deleted apache but I am still getting the "It works!" in my browser.
[07:10] <sarnold> Syria: if you force a refresh iwth something like shift+reload, is it still there?
[07:10] <sarnold> Syria: if you test with a simpler browser such as w3m or lynx is it still there?
[07:10] <Syria> sarnold: Yes, Also using another browser did not change anything.
[07:10] <sarnold> Syria: does netstat -lnp | grep :80 show apache running?
[07:12] <Syria> sarnold: http://paste.ubuntu.com/5604194/
[07:13] <sarnold> Syria: re-run with 'sudo', the last line of output sure looks like a webserver
[07:13] <Syria> sarnold: tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      3798/apache2
[07:13] <sarnold> yay :)
[07:13] <sarnold> Syria: kill 3798
[07:14] <sarnold> Syria: how did you delete apache? I'm surprised a process is left over...
[07:15] <Syria> sarnold:  apt-get remove the sudo update-rc.d -f apache2 remove
[07:15] <Syria> *then
[07:15] <sarnold> Syria: I must admit I don't delete apache often :) but I would have expected that to kill apache.
[07:16] <sarnold> okay bed time :)
[07:17] <Syria> sarnold:  sudo apt-get remove apache2 >>> is thic correct ?
[07:29] <Syria> Anybody knows why I am still getting the "It works!" after removing apache2 please?
[07:31] <Syria> Lost the connection.
[07:32] <Syria> I was asking about apache, After removing apache2 I am still getting the "It works!"
[09:11] <jamespage> rbasak, around? I have a favor to ask
[09:12] <rbasak> jamespage: morning! What's up?
[09:12] <histo> favors??
[09:12] <jamespage> rbasak, morning!
[09:12] <jamespage> rbasak, so..... mongodb
[09:12]  * rbasak grabs an ARM machine from the lab :-P
[09:12] <jamespage> rbasak, I have a couple of issues on ARM and I'm well out of my depth and I wondered whether you had time to take a look?
[09:13] <rbasak> Sure!
[09:13] <jamespage> rbasak, OK - so issue one relates to the arm-support patch that is in the current raring packaging
[09:13] <jamespage> rbasak, I picked this from the fedora-arm ML
[09:14] <jamespage> rbasak, but on arm mongod just spins on startup in atomic_intrinsics_gcc.h
[09:14] <jamespage> rbasak, so I'm not convinced its really working right :-)
[09:14] <jamespage> i had a hack using gcc __sync builtins and got it to start but then hit issue 2)
[09:15] <jamespage> rbasak, issue 2 is I think something todo with alignment
[09:15] <psivaa> hallyn: ack, thanks
[09:16]  * rbasak looks
[09:16] <jamespage> rbasak, you can see this by running the mongo command
[09:16] <jamespage> it SIGBUS's straight away
[09:17] <jamespage> rbasak, src/mongo/bson/bsonelement.h:581
[09:17] <jamespage> rbasak, OH - btw I enabled ddeb's for mongodb - thats already inthe archive
[10:14] <koolhead17> Daviey: jamespage around?
[10:15] <koolhead17> needed confirmation if i should cloudrepo 4 swift installation as well if am trying folsom on precise?
[10:16] <Daviey> koolhead17: you should.
[10:16] <Daviey> cloud archive
[10:17] <koolhead17> Daviey: thanks!!
[10:41] <NaGeL_Work> !webmin
[10:42] <SystemParadox> morning all. Please can someone help me out with environment variables and cron. I'm fairly experienced, but cannot figure this out. Even if crontab has SHELL=/bin/bash and is executing a script with #!/bin/bash, the PATH is *still* wrong- it's not using /etc/environment, but then I can't find any documentation for this file either- is this bash or ubuntu specific? Thanks
[10:43] <NaGeL_Work> is there any alternatives for webmin?
[10:43] <NaGeL_Work> that is supported by ubuntu?
[10:49] <NaGeL_Work> or is there any guide how to set up your web server? and secure it?
[11:48] <NaGeL_Work> is virtualmin supported by ubuntu-server?
[11:48] <jpds> N: Unable to locate package virtualmin
[11:48] <jpds> No.
[11:49] <NaGeL_Work> any alternatives to that? or webmin?
[11:49] <jpds> !webmin | NaGeL_Work
[11:50] <NaGeL_Work> Alternatives?
[11:50] <Patrickdk> ssh?
[11:51] <NaGeL_Work> something with a GUI. shh isa back up for me due that i don't really trust my command line knowledge
[11:54] <koolhead17> Daviey: planning to give grizzly try from tomorrow :D
[11:54] <koolhead17> deb http://ubuntu-cloud.archive.canonical.com/ubuntu precise-updates/grizzly main  <-- this is what i will use
[12:02] <Diegonat> hi guys! I have got a huge problem. I have a luks partition /dev/mapper/archive-vmbackup but when I try to close it, "cryptsetup --verbose luksClose /dev/mapper/archive-vmbackup"
[12:02] <Diegonat> "Command failed with code 19: No such device" I have got this output... Do you have any idea?
[12:28] <zenzio> "I’m glad Sony decided to go with 8gb RAM because it means that the PS4 will out-power most PC’s for years to come."... say what?
[12:29] <psivaa> jamespage: just seen this one time post install failure with postgresql raring smoke tests: https://jenkins.qa.ubuntu.com/view/Raring/view/Smoke%20Testing/job/raring-server-amd64-smoke-postgresql-server/115/
[12:30] <psivaa> jamespage: not sure what's causing and it only happened once. dont see much difference from the previous passing install logs.
[12:30] <zenzio> er... that was the wrong channel, sorry
[12:31] <psivaa> jamespage: it would help if you could shed some light when you have some time please. thanks
[12:33] <NaGeL_Work> I goat question which would be better configurable/ more useful : zend server CE or plain apache2 and installing APC to it(and other useful stuffs)
[12:47] <rbasak> jamespage: looks like there are already arrangements in place to avoid the unaligned access for the double on ARM. Evidently it's not working though. I'm still looking at it.
[12:48] <jamespage> rbasak, yeah - some of that came from work in ubuntu for the 2.0.x series
[13:13] <jamespage> yolanda, any chance you can take a look as to why ceilometer is ftbfs in the lab please?
[13:13] <yolanda> sure
[13:39] <hallyn> stgraber: 1. patch hit libvirt m-l for userns support for libvirt-lxc
[13:39] <hallyn> stgraber: 2. i intend to push my hacky fix for lxc-execute to raring today, so that the utah tests can pass (or go on to the next failure :)
[13:47] <stgraber> hallyn: ok. I'm about to send an e-mail to lxc-devel saying that the rc1 pull request will go out on Wednesday
[13:49] <yolanda> jamespage, i'm having a problem with README.md in docs
[13:50] <yolanda> looking at git it's there but i have this error: dh_installdocs: cp -a README.md debian/python-ceilometer/usr/share/doc/python-ceilometer returned exit code 1
[13:57] <Xaro> Hey
[13:59] <jamespage> yolanda, OK - I'll take a look
[13:59] <jamespage> yolanda, in the meantime - https://code.launchpad.net/~james-page/keystone/fixup-test-overrides/+merge/152680
[13:59] <yolanda> i tried to move file from docs to ceilometer-common.docs but still same problem, let me send you the MP anyway because i fixed other problems
[14:04] <yolanda> jamespage, found that ceilometer isn't following the other packages in .docs, i'll take a look at nova, glance... and try to do the same
[14:05] <jamespage> yolanda, hmm
[14:05] <jamespage> I don't see that issue
[14:06] <yolanda> apart from removing patches, it builds for you?
[14:06] <yolanda> if i remove docs file it works, but as soon as i add it, i have this problem
[14:06] <jamespage> yolanda, yep
[14:06] <yolanda> that fails for me :(
[14:07] <jamespage> yolanda, all I did was rm the patches folder
[14:07] <yolanda> yes, i did the same
[14:07] <yolanda> but it doesn't build due to that docs problem
[14:07] <yolanda> how do you generate the source? i pull from git, then i do an sdist, and use that
[14:09] <jamespage> yolanda, I have a hack
[14:09] <jamespage> yolanda, http://paste.ubuntu.com/5604865/
[14:09] <yolanda> maybe is the way that i do that, that it causes problem? it always worked for me
[14:13] <yolanda> jamespage, but it shouldn't be the same as doing a python sdist?
[14:13] <jamespage> yolanda, I think so yes
[14:13] <yolanda> let me try with that script anyway
[14:13] <yolanda> because i cannot build
[14:17] <yolanda> jamespage, seems that code shouldn't be the same, i can build with your tool to download source, and not with sdist
[14:18] <yolanda> i'll submit the MP
[14:19] <Quest> hi, I have question about google serach
[14:20] <eutheria> i've setup ubuntu to do some routing for me, i have a weird issue where the traceroute going in one direction takes a long time with some timeouts
[14:20] <Quest> my domain is new, I have a link to sub.mysite.com  in the mysite.com/index.html. mysite.com is visible in google search but sub.mysite.com is not(not by keywords of its contents nor by its name . what could be the reason?
[14:22] <YamakasY2> guys is it doable to install a server with 5GB for usr and var ?
[14:27] <yolanda> jamespage https://code.launchpad.net/~yolanda.robla/ceilometer/grizzly/+merge/152686
[14:28] <jamespage> yolanda, did you mean to leave that change in ceilometer-common.install
[14:33] <yolanda> jamespage, no, sorry, it was due to my problems with building, it was also giving problems with that file
[14:33] <yolanda> i'll correct it
[14:34] <yolanda> i didn't commit the file :(
[14:39] <yolanda> jamespage, recheck
[14:57] <YamakasY2> guys, do you seperate var and usr ?
[15:07] <Quest>  disconnected..
[15:19] <Quest>  so i do need site maps and robots.txt ?  just linking pages with each other by hyperlinks is not suffice for subdomains?
[15:29] <jamespage> yolanda, conditional approve - please update the changelog and merge
[15:31] <yolanda> ok
[15:41] <YamakasY2> I think it's still wise to seperare /usr and /var but what should be the size ? wil 5GB per partition do ?
[15:51] <stgraber> hallyn: can you review my lxc-lenny removal patch on lxc-devel?
[15:54] <hallyn> didn't i ack that already?
[15:56] <hallyn> guess not
[15:58] <stgraber> hallyn: not that I can remember
[15:58] <stgraber> hallyn: I also just sent another batch of 3 changes. Adding get_version to the API and tweaking the python binding. I think that's mostly it for changes I have for rc1 here.
[15:59] <hallyn> stgraber: should we take the kmsg patch?  (if so we should add that option to lxc-ubuntu template's boilerplate)
[16:00] <hallyn> stgraber: got one of your 3 patches so far :)
[16:01] <hallyn> ah there they are
[16:01] <stgraber> hallyn: oh, I meant to comment on this one. I'm fine with it but I'd prefer the default to be "1".
[16:02] <hallyn> stgraber: agreed :)
[16:02] <stgraber> hallyn: the problem with changing the templates is that it won't upgrade existing containers
[16:02] <hallyn> true
[16:02] <hallyn> stgraber: of course there is one argument in favor of getting rid of kmsg handling -
[16:03] <hallyn> it'll annoy users so they ask for proper syslog ns :)
[16:04] <hallyn> stgraber: patch 2 - did you forget to add version.c to the index?
[16:04] <hallyn> can you resend?
[16:05] <stgraber> hallyn: version.c is already upstream, it was just missing version.h
[16:05] <stgraber> (and wasn't really used as far as I can see)
[16:06] <hallyn> d'oh
[16:06] <hallyn> i never knew that was there :(
[16:06] <hallyn> uh, i meant :)
[16:16] <syncsys_> http://stackoverflow.com/questions/15342665/google-seo-indexing-subdomains
[16:25] <jamespage> yolanda, review pretty please :-) https://code.launchpad.net/~james-page/quantum/iputils-arping/+merge/152716
[16:26] <yolanda> done
[16:36] <syncsys_> can anyone answer http://stackoverflow.com/questions/15342665/google-seo-indexing-subdomains
[16:51] <lunaphyte_> hi.  my network interfaces are being named things like emo, p1p1, p1p2, etc, instead of eth0, eth1, etc.  how can i turn this behavior off?
[16:51] <lunaphyte_> *em0
[16:52] <sarnold> lunaphyte_: I think the answer is to fiddle with udev's 70-persistent-net.rules file
[16:53] <sarnold> (see https://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Deployment_Guide/appe-Consistent_Network_Device_Naming.html for details about the change)
[16:53] <melmoth> http://mickvaites.com/2009/06/ubuntu-changing-network-device-id-udevrules-d/ (looking for "ubuntu udev network interface name" )
[16:53] <lunaphyte_> great, thank you.  i had found a number of references to udev, but not those two
[16:54] <lunaphyte_> it's sort of odd too though - i have installed 12.10 quite a few times on a number of different computers, and this is the first time i've seen it use this different device nomenclature
[16:56] <sarnold> hehe, I haven't seen it yet myself either..
[16:56] <tedski> yeah, i haven't seen driver-based naming on ubuntu... are you sure you didn't install centos6 :P
[17:06] <acalvo> Hello
[17:07] <acalvo> Preseeding seems to fail to load the preseed file using the alternate install cd and setting up isolinux to load it from the cdrom
[17:07] <acalvo> however, passing parameters (such as locales) do work
[17:07] <acalvo> I'm not using ubiquity as an unattended installation may not have a graphical frontend runnig
[17:14] <xnox> acalvo: too many negatives. Are you preseeding? Which image are you using? (pxe, mini, server, alternate, desktop)? And which release?
[17:16] <acalvo> I'm using an LTS (12.04) alternate image (x86_64)
[17:28] <ScottK> stgraber: I think we're about at the time where migration freezes should go in place.  Since you volunteered to take care of that ....  If you look in the bzr history for the block I put up for Alpha 1/2, please just use that one again.
[17:28] <ScottK> Oops.
[17:28] <ScottK> Wrong channel.
[17:32] <jamespage> smoser, utlemming: are there any know issues with raring with regards to virtualized performance?
[17:32] <jamespage> (before I go dig into something I saw end of last week)
[17:32] <smoser> jamespage, not that i'm aware of.
[17:32] <dragoonis> command not found "do-release-upgrade"
[17:32] <smoser> what does "virtualized performance" mean ?
[17:32] <dragoonis> I'm on Ubuntu 10.10, trying to get up to 12.x
[17:33] <jamespage> smoser, so I'm sure I saw a much higher host load average when deploying openstack-ha on virtual maas when compared to precise
[17:33] <jamespage> smoser, like I said I need to dig deeper I think
[17:33] <dragoonis> got it.
[17:48] <jamespage> yolanda, oh great - ceilometer has grown hbase support - https://github.com/wbolster/happybase
[17:49] <yolanda> i saw some comments about that in the channel last week
[17:49] <yolanda> they have already deployed it?
[17:50] <jamespage> yolanda, its in the codebase; we might have to disable that bit for grizzly
[17:50] <jamespage> yolanda, its a bit late to be uploading new packages and happybase is not packaged
[17:50] <yolanda> ok, i can take a look
[17:51] <yolanda> we are recommending mongo in our package
[17:52] <jamespage> yolanda, yes
[17:52] <jamespage> hbase is not even packaged
[17:52] <jamespage> charmed - yes
[17:52] <jamespage> packaged - no
[17:53] <jamespage> yolanda, (found that whilst trying your charm btw :-))
[17:54] <yolanda> ok, i'll take a look tomorrow morning
[17:54] <yolanda> how is my charm anyway?  :)
[18:12]  * tgm4883 scowls at Daviey 
[18:12] <Daviey> tgm4883: ?
[18:12] <tgm4883> http://mythbuntu.org/
[18:13] <tgm4883> Daviey, did you try adding the DNS entries as provided by Google?
[18:13] <Daviey> hmm
[18:16] <Daviey> -> taken to a different channel
[18:41] <Vrga> Hello folks, question for anyone who knows, under which user should the /var/www and folders therein be? www-data or?
[18:41] <Vrga> (apache2 related question)
[18:43] <RoyK> usually root
[18:44] <RoyK> you don't want www-data to be able to write to stuff like php files
[18:44] <RoyK> if you need uploads, create a separate dir writable by www-data
[18:44] <RoyK> that is - no need for it to be owned by root, just not www-data
[18:46] <Vrga> ah, ok.
[18:50] <RoyK> Vrga: thing is, if you have a php script that is somewhat flakey, write permission for www-data may allow the php scripts to be rewritten by code injection
[18:51] <Vrga> currently, i'm trying to figure out why my virtual hosts are being iffy
[18:51] <Vrga> thought it was iffy permissions, but its something else i think...
[18:51] <RoyK> what's iffy about them?
[18:51] <sarnold> Vrga: check your logs for details?
[18:52] <RoyK> if it's successful code injection, it probably won't be logged (at least not if using POST)
[18:52] <Vrga> RoyK, too long to explain at the moment, need to collect all the details and thoughts on the whole issue. And if it were as trivial as tailing logs, i wouldnt be ere :D
[18:52] <sarnold> Vrga: it's all too often overlooked :(
[18:53] <sarnold> RoyK: too true :)
[18:53] <Vrga> i know, i know :D
[19:06] <Vrga> btw, what about wsgi scripts? do they need to be under the www-data user or root user?
[19:08] <lunaphyte_> sarnold, tedski: yeah, same here.  what's also a little odd is that /etc/udev/rules.d/70-persistent-net.rules is not being regenerated when i reboot.
[19:09] <sarnold> lunaphyte_: I _think_ it is rebuilt only on special occasions, and probably leaves alone things you have configured yourself. (i've forgotten the conditions now, and I'm not entirely sure where to look...)
[19:10] <lunaphyte_> hmm, ok.
[19:28] <lunaphyte_> aha!  i found the culprit :)
[19:28] <lunaphyte_> http://packages.ubuntu.com/quantal/biosdevname
[19:29] <lunaphyte_> apt-get --purge autoremove biosdevname makes it go back to the "traditional" behavior
[19:30] <lunaphyte_> https://lists.ubuntu.com/archives/ubuntu-devel/2012-January/034687.html
[19:30] <lunaphyte_> this particular computer happens to be an optiplex 990.
[19:30] <lunaphyte_> so although obviously not a poweredge server, certainly a dell.
[19:31] <lunaphyte_> and 70-persistent-net.rules was regenerated this time.
[19:34] <lunaphyte_> it's also mildly interesting that the disk this computer is booting from came out an optiplex 760, where i did not see this behavior.
[19:44] <stgraber> hallyn: do you agree with Dwight's proposed change in syntax for the userns uid mapping? I don't mind changing it now, but ideally it should stay that way from now on ;)
[20:12] <lunitik> lunaphyte_, I really recommend leaving it around, it actually makes more sense than the traditional naming, especially if you have multiple nics or something like that
[20:14] <lunitik> lunaphyte_, Usually, for instance, whichever nic comes up first is eth0... in the new naming, it is by physical location, so you always know which is which
[20:19] <lunaphyte_> lunitik: yeah, i can appreciate that.  it's an interesting concept, and one i'm not empirically discarding, but for the moment, i just want the familiar behavior.
[20:20] <lunaphyte_> abrupt changes in device nomenclature have the obviously spectacular potential to break a *ton* of stuff.
[20:21] <lunaphyte_> i can't even begin to list the litany of things which expect ethernet interfaces to be ethn
[20:21] <lunitik> lunaphyte_, Well, theoretically, you would take steps to handle the change within the update process.
[20:22] <lunaphyte_> oh, of course.  ultimately, it's a non issue.  but that doesn't absolve the identification mechanism.
[20:23] <lunitik> lunaphyte_, I am not actually sure why anything would expect ethn... it isn't like its an actual /dev entry, but I'm not familiar with how that works from a programming perspective.
[20:23] <lunitik> lunaphyte_, Seems the change to sd from hd would have caused more problems, but I don't recall anything
[20:24] <lunaphyte_> fair point.
[20:24] <lunaphyte_> it did cause human confusion, of course, just as this did for me at first :)
[20:25] <lunitik> lunaphyte_, Oh, I looked into biosdevname exactly because I was confused about it, too, but it really does make a lot of sense
[20:28] <lunaphyte_> a lot of that sort of thing that would expect ethn is truly just poorly written software which is making assumptions that shouldn't be made, but it's been such an axiomatic characteristic of linux for so long, it's almost become gospel.
[20:31] <lunaphyte_> yeah, i can appreciate the theory, for sure.  i just have to give some thought to my opinion of the practical implications of the implementation.
[20:33] <hallyn> stgraber: haven't seen that email but if it's what he's mentioned to me before then yes
[20:33] <stgraber> hallyn: [lxc-devel] [PATCH] make [ug]id map ordering consistent with /proc/<nr>/[ug]id_map
[20:34] <hallyn> yup that sounds good
[20:34] <stgraber> ok, I'll pull that in staging then. The diff itself looks good to me. I just wanted to make sure you agreed with the syntax change.
[20:37] <hallyn> stgraber: for the record i'll also ack it on the list
[20:37] <stgraber> hallyn: ok
[20:42] <hallyn> stgraber: for his other patch, i'm checking right nwo - i'm still not sure he's correct about the kernel limtiation
[20:48] <daguz> I cannot seem to get a Ubuntu server to install as a paravirtualized guest under xen.  Does anyone have any hints?
[20:48] <daguz> I never see a /proc/sys/xen directory.  So that is why I assume it is not PV.
[20:49] <daguz> Even though I'm trying to do quantal I'm following (basically) these instructions: http://www.mmacleod.ca/blog/2011/05/ubuntu-natty-narwhal-and-xen/
[21:00] <hallyn> well whaddyaknow, i'm just plum wrong
[21:06] <sarnold> o_O
[21:06] <sarnold> check again
[23:24] <wpbiggs> hi, can someone tell me if i can use bind in ubuntu server to host my own domain name, or if i misunderstand the technology?
[23:26] <sarnold> wpbiggs: you sure can, see this guide for one way to do that: https://help.ubuntu.com/12.04/serverguide/dns.html
[23:28] <wpbiggs> thanks for your reply, so if i follow this guide, i would not have to use a service like godaddy or anything else to regster the domain?
[23:28] <sarnold> wpbiggs: that is different :)
[23:29] <wpbiggs> that's what i was afraid of is there a way to do that myself?
[23:29] <sarnold> wpbiggs: you need a domain registrar such as godaddy or gandi or namecheap; they will typically require you to have two or more nameservers registered with them (aka "in glue") and those nameservers do the DNS...
[23:31] <wpbiggs> oh, thanks, i had tried to host my domain before a few months ago, but i guess that was the step that i missed.